Importance of SoilContaminantSurfactant ... - ACS Publications

Chapter 16

Importance of Soil—Contaminant—Surfactant Interactions for In Situ Soil Washing 1

2,3

1

R. C. Chawla , C. Porzucek , J . N . Cannon , and J . H . Johnson, Jr. 1

1

School of Engineering, Howard University, Washington, DC 20059 Los Alamos National Laboratory, Los Alamos, N M 87545

Downloaded by FUDAN UNIV on March 15, 2017 | http://pubs.acs.org Publication Date: August 2, 1991 | doi: 10.1021/bk-1991-0468.ch016

2

Organic pollutants can bind to soil matter by various types of phys ical and chemical interactions—hydrophobic bonding, hydrogen bonding, ion exchange, charge transfer, etc. They enter the en vironment from spills, leaks, leaching and other sources. They are emplaced in soils through capillary action or by the physical and chemical forces of sorption. The displacement and recovery of or ganic pollutants by in situ soil washing may use aqueous surfac tant solutions. In a surfactant-assisted soil-washing operation, soil -contaminant-surfactant interactions, and the surrounding soil en vironment, all have major effects on the pollutant mobilities. The various soil-contaminant-surfactant properties, and their effects on the removal of organics, are reviewed. The interactions between soil/aquifer systems, contaminants, and surfactants are discussed in the context of the fate of contaminants in the subsurface system. Remedial action plans for cleanup of contaminated soils generally fall into two categories. They are either above ground or in situ activities. The above ground techniques require excavation of the contaminated zone followed by some treatment such as bioremediation, soil washing, or incineration. The in situ techniques are used on contaminated sites that are either very large, very inaccessible, or both. One example of an inaccessible site is a gasoline spill that occupies the soil beneath an airplane hangar. Another example is a leak from a pipe that is buried 100 feet below the ground. For many inaccessible sites, in situ cleanups are the only practical treat ment methods. However, in situ cleanups are more difficult than above ground cleanups because it is not possible to have complete control over critical soil remediation process parameters (e.g., temperature, pH, and the spatialflowof fluids). Consequently, in situ cleanups are riskier than above ground actions. 3

Current address: Dow Chemical Texas Operations, Freeport, TX 77541 0097-6156/91/0468-0316$07.50/0 © 1991 American Chemical Society Tedder and Pohland; Emerging Technologies in Hazardous Waste Management II ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

16. CHAWLAETAL.

Soil-Contaminant-Surfactant Interactions

317

The Chemical Countermeasures Program, sponsored by U.S. Government agencies like the Environmental Protection Agency (EPA), the Department of Defense, and the Department of Energy, is designed to evaluate in situ reme diation processes for the cleanup of contaminated soils (1,2). Sites that are candidates for these cleanup methods should have the following characteris tics: 3

1. contaminants are spread over large volumes (e.g., 100 to 100,000 m at a depth of 1 to 10 m), 2. contaminant concentrations are low (i.e., less than 10,000 ppm), 3. contaminants can be removed or immobilized by aqueous chemical solu tions injected into the contaminated site, and 4. the hydraulic conductivity of the contaminated soil is greater than 10~ cm/s.


4

Two types of in situ cleanup technologies have been tested. First, acids and chelating agents have been evaluated for the in situ remediation of soils con taining hazardous heavy metals. Secondly, surfactant washing techniques have been considered for the in situ remediation of soils contaminated by hazardous organics. Chawla et al. (3) have critically reviewed past work on surfactant-assisted, in situ soil washing for the removal of hazardous organics. Laboratory, pilot, andfield-scalestudies have been conducted to investigate and evaluate this re medial action technique. Initial laboratory studies were conducted at the Texas Research Institute (TRI) (4,5). They showed the effectiveness of this technique in recovering gasoline from contaminated sands. For example, one set of ex periments showed that surfactants could be used to mobilize residual gasoline in the vadose zone. Once mobilized, the gasoline migrated downward to the water table where it was recovered. A second set of experiments tested differ ent methods for application of the surfactant solution. The best case results from both types of experiments suggested that about 80% recovery of residual gasoline from the sands was possible. Surfactant-assisted, in situ soil-washing studies have also been conducted by the Science Applications International Corporation, Inc. (SAIC) (1,2). These studies showed the effectiveness of this technique in removing polychlorinated biphenyls, Murban crude oil, and phenols from doped soil samples using mix tures of nonionic, biodegradable surfactants. Results from core flood experi ments suggested that about 68-88% of the pollutants could be removed. After these successful lab experiments, SAIC conducted a second set of experiments using contaminated soil from a fire fighting area of an Air National Guard Base (6). The soil was contaminated with petroleum derived hydrocarbons (JP-4 jet fuel), chlorinated solvents (1,1,1-trichloroethane, dichloromethane, trichloroethylene, and chloroform), and afire-fightingfoam. In these tests they were able to remove 90-95% of all contaminants (i.e., aliphatic, aromatic, and unresolved hydrocarbons).

Tedder and Pohland; Emerging Technologies in Hazardous Waste Management II ACS Symposium Series; American Chemical Society: Washington, DC, 1991.


318

EMERGING TECHNOLOGIES IN HAZARDOUS WASTE MANAGEMENT II

Following these lab studies, a field test was conducted by Mason & HangerSilas Mason Company at the same Air National Guard Base (7,8). Aqueous surfactant solutions were applied to the contaminated soil surface and allowed to percolate downward. This type of surfactant solution application should have drained contaminants from the vadose zone down toward the water ta ble, but results showed essentially no contaminant movement. The failure was attributed to the development of preferential flow paths that bypassed contam inated areas of the soil. These preferential flow paths developed after many years of rainfall accompanied by a redistribution offine-sized,soil particles. More recent studies were performed to evaluate this technology for reme diating soils at a railroad-tie treating plant in Laramie, Wyoming. Soil contam inants included creosote oil, polynuclear aromatics, and pentachlorophenol. Since this site was severely contaminated, a multistep remediation process was adopted. Initially, contaminants were pumped directly out of the ground, simi lar to a primary oil recovery operation. This action was followed by a waterflood for secondary contaminant recovery. Then a surfactant flood, designed by M T A Remedial Resources, Inc., was implemented (9). The formulation consisted of a surfactant, alkali, and polymer mixture. While their laboratory studies had shown greater than 98% recovery of creosote from contaminated soils with this surfactant wash, pilot tests at the railroad-tie treating facility only yielded a 67% reduction in contaminant concentrations. However, the overall recovery from primary pumping, secondary waterflooding, and the tertiary surfactant flood was 94%. Another recent implementation of this technology occurred at an industrial site in Florida which was contaminated with a viscous oil (100 cp at 20 °C). The oil existed as a free layer (0.01 to 1.5 feet thick)floatingon the water table over a 40,000 ft area. Their contaminant layer was separated from the drinking water supply aquifer by only a semipermeable silt layer. Again, M T A Remedial Resources designed a surfactant system for an in situ soil wash (10). Due to the location of the contaminant and its proximity to the drinking water aquifer, the surfactant system consisted of "chemicals used in water treatment and as food additives" and a "biodegradable polymer". Field results showed that these washing solutions were able to remove up to 75% of the trapped oil. In 1988, a joint study was initiated between Howard University and Los Alamos National Laboratory to determine the feasibility of using surfactants for in situ remediation of contaminated soils and aquifers. Thefirststep of the study was to identify the mechanisms behind the transport of contaminants in soils and aquifers. More specifically, the mechanisms of interaction between contaminants, soils, and ground water were identified. This study was followed by an investigation into the effects of surfactants on each mechanism. The re sults were used to explain successes and failures of previous laboratory and pilot-scale field studies of surfactant-enhanced, in situ soil washing (3,11). The design of effective in situ remedial action plans requires a thorough 2



16.

CHAWLA ET AL.


319

understanding of physical chemistry and hydrogeology. The most common in situ remedial action plans for ground water cleanup use the pump-and-treat philosophy. This philosophy treats the contamination problem like a "black box." Liquid is pumped from the aquifer and treated on the surface. Pumping is continued until an acceptable level of aquifer contamination is reached. Recent studies show, however, that contaminants can take decades to dissolve in the ground water (11,12). This makes pump-and-treat plans slow and expensive. In order to design effective in situ remedial action plans, it is necessary to understand the interactions between contaminants, soil, and ground water. Here we describe the relevant physical properties of these components and their interactions with each other and with aqueous surfactant solutions. Ex planations of sorption/desorption, capillary forces, and mass-transfer limited solubilization are given along with the influence of surfactants on each. Other mechanisms that influence the transport and retardation of contaminants in soil and aquifer systems are also discussed. Multiple processes are responsi ble for the transport of contaminants in a soil and aquifer system. The most effective remedial action plans will be those whose designs are based on an understanding of these fundamental transport phenomena.

The Contamination Problem In order to define the mechanisms of contaminant transport, it is necessary to develop a physical picture of a contaminant discharge into the environment. For example, spills or releases of nonaqueous phase liquids (NAPLs) come from sources such as leaky storage tanks and pipes, and spills on the ground. The typical movement of contaminants following a release is illustrated in Fig ure 1. As the contaminants move, they are acted upon by various forces in the soil. Initially a contaminant, referred to as a NAPL, moves downward through the vadose zone because of the influence of gravity. Some of it is held trapped in the vadose zone by capillary forces. This N A P L is called a "residual" and it exists as small blobs and pendular rings in the soil pore structure (13). Ex perimentally, Hoag and Marley (14) found residual gasoline concentrations in the vadose zone to be about 14% in coarse sands with average particle diam eters around 2.2-mm, and as high as 60% in fine sands with average particle diameters about 0.26-mm. The N A P L continues to migrate downward through the vadose zone until it encounters the water table, also shown in Figure 1. At this point, a lighter-thanwater N A P L will spread out and form a lens on the water table. On the other hand, a heavier-than-water N A P L will continue to migrate downward through the aquifer until it hits an underlying rock layer and forms a N A P L pool. In both cases, N A P L will dissolve into the aquifer and create a contaminated ground water supply.


320



NAPL SPILL

NAPLS DISSOLVED IN GROUND WATER

Figure 1. Migration of a Contaminant Spill



16. CHAWLA ET AL.

SoU-Contaminant-Surfactant Interactions

321

Due to naturalfluctuationsin aquifer recharge and discharge or pumping, the height of the water table willfluctuatethroughout the year. This results in a distribution of trapped, lighter-than-water, N A P L lenses at different depths in the aquifer and vadose zone. The N A P L lenses trapped below the water table were emplaced there when the water table fell. As the water table rose again, the buoyancy forces exerted on the N A P L phase were insufficient to overcome the capillary forces that held the N A P L trapped in the pores. The end result is a N A P L trapped below the water table. A similar situation occurred to emplace NAPLs in the vadose zone. Here, the water table fell and never recovered back to its original height. In this case, the capillary forces were larger than the gravity forces and the N A P L became trapped. In addition to N A P L dissolution into the ground water, it will migrate away from the original contamination zone due to the natural hydraulic gradient of the aquifer. These NAPLs will be pushed along as a separate phase by the viscous forces of the ground water flow. However, they may encounter soil pores that exert sufficient capillary forces to entrap them. Such trapped NAPLs will no longer exist as lenses, but instead as ganglia and blobs. Wilson and Conrad (13) estimated residual N A P L saturations of 10% to 50% in saturated (aqueous) soil systems. Residual NAPLs also pose a long-term contamination problem because they slowly dissolve into the ground water. From this description of N A P L movement, it is clear that NAPLs can exist in any of several forms in a soil/aquifer system. They may occur as a residual phase, or as trapped lenses, or as dissolved species. The exact disposition of a particular contaminant depends on the specific physical properties of a surfac tant, contaminant, and soil. It also depends on the interactions between these species as described below.

Soil/Aquifer-Contaminant-Surfactant Interactions In a surfactant-enhanced, in situ soil-washing operation, surfactant, contami nant, and the soil/aquifer system are the three interacting components. Here the term soil/aquifer system is used to mean soil in general. This includes the saturated as well as unsaturated (vadose) zone. Each component interacts with the other two as illustrated by the triangular diagram of Figure 2. In this dia gram, each of the three components occupies an apex. The interactions be tween any two components are listed along the line connecting the two respec tive apexes. These component-component interactions can be physical, chem ical or biological in nature. Soil/Aquifer System Properties. Properties of the soil/aquifer system that are important in the transport of contaminant and surfactants are shown in Fig-



322

E M E R G I N G TECHNOLOGIES I N HAZARDOUS WASTE M A N A G E M E N T II

[SURFACTANTI HYDRAULIC GRADIENT

N A P L SOLUBILIZATION CHEMICAL REACTION

SORPTION D E C R E A S E IN I FT BIODEGRADATION

DESORPTION

CONTAMINANT

I SOIL / A Q U I F E R HYDRAULIC GRADIENT CAPILLARY FORCES BUOYANCY SOLUBILITY SORPTION BIODEGRADATION VOLATILIZATION PLANT UPTAKE CHEMICAL REACTION

Figure 2. Interaction Triangle for Soil/Aquifer-Contaminant-Surfactant



16. CHAWLA ET A L

Soil—Contaminant—Surfactant Interactions

323

ure 3. Temperature is important because surfactant and contaminant solubil ities are functions of temperature. Species sorption properties are also tem perature dependent. Therefore, it is important to know the temperature of the soil/aquifer system before a surfactant-enhanced soil washing system is de signed. Soil/aquifer system pH is also a critical variable. The pH will determine the surface charge on any minerals present in the soil/aquifer system. It will also af fect the configuration of any humic materials (e.g., coiled or uncoiled) (15,16). Both of these changes will in turn quantitatively affect the amount of surfactant or contaminant sorbed. In addition to pH, organic carbon content of the soil/aquifer system has been shown to correlate well with the amount of contaminant sorbed onto the soils (17-20). Therefore, it is necessary to have an idea of the location and amount of organic carbon. The particle size distribution of the soil/aquifer system is also important. Soils with larger average particle sizes, say greater than 62-μπι, are usually sandy and have little or no sorptive capacity. On the other hand, soils with smaller particles (e.g., average diameters less than 2-μπι), usually contain more clay minerals. These clay minerals have a sorptive capacity for contaminants. Also, soils with small particles tend to be less permeable to fluid flow. In soil/aquifer systems that are saturated with water, the porosity, perme ability, and pressure all affect the flow field of the ground water. Ground water tends to follow paths that lead through more permeable sections of the soil. It also tends to follow paths through soil sections that have higher porosity. F i nally, ground water will travel the fastest from areas of high pressure to areas of low pressure. Therefore, it is helpful to have an idea of the permeability, porosity, and pressure distribution in a soil/aquifer system before designing an injection and recovery well system for remediation. Using this map it may be possible to design a system that maximizes the flow through the contaminated zone and also recovers fluids that have been swept through this zone. Finally, the moisture content of the soil/aquifer system is a critical variable. For remedial action design purposes it is necessary to know if the contamina tion problem exists in the vadose zone (low moisture content) or the aquifer (saturated zone). Flow in the vadose zone is more unpredictable when com pared to the saturated zone (aquifer) because of the presence of air. On the other hand, flow in the aquifer can be predicted well if one has an idea of the permeability and pressure distribution in the aquifer. Contaminant Properties. In order to remove or clean up a contaminant one must define the contaminant and its physical properties. These properties will give some idea as to the mobility of the contaminant and how difficult it will be to remove from the soil/aquifer system. Important contaminant properties are shown in Figure 4.


324


-TEMPERATURE


-PH - HUMIC & ORGANIC CARBON CONTENT - PARTICLE SIZES (SANDS, CLAYS) - POROSITY - PERMEABILITY - PRESSURE DISTRIBUTION - MOISTURE CONTENT

Figure 3. Physical Properties of Soil/Aquifer

J CONTAMINANT

- AQUEOUS SOLUBILITY - HYDROPHOBIC / HYDROPHILIC - ORGANIC / INORGANIC - POLAR/NONPOLAR - DENSITY - VISCOSITY

Figure 4. Physical Properties of Contaminant



16. CHAWLA ET AL.

SoU-Contaminant-Surfactant Interactions

325

The aqueous solubility of the contaminant will determine if the contaminant is a major threat to the ground water. A highly water-soluble contaminant will spread into the ground water and increase the zone of contamination much faster than a contaminant that is only slightly water soluble. On the other hand, a slightly water-soluble contaminant poses a long-term threat to the ground water because it will take longer to dissolve. Nonpolar contaminants are those with little or no net dipole moment. Such contaminants are hydrophobic and hence only slightly water soluble. On the other hand, polar contaminants have a net dipole moment. These contami nants are hydrophilic and hence highly water soluble. The transport properties of polar and nonpolar contaminants in the soil/aquifer system will be different. It is also important to know if the contaminant is organic or inorganic since the remediation schemes differ for these two categories. In addition, the den sity of the contaminant must be known. As mentioned previously, a contami nant that is denser than water will lie predominantly beneath the aquifer while a less dense contaminant will be found floating on top. Finally, it is important to know the contaminant viscosity. If the viscosity is too high, it may not be possible to use a hydraulic gradient for site remediation. Surfactant Properties. Surfactant properties that are important in the re mediation process are shown in Figure 5. A surfactant must increase N A P L solubilization in the aqueous phase and reduce interfacial tension between the aqueous and N A P L phases. In general, surfactants can be cationic, anionic, or nonionic. Since ionic surfactants have a tendency toflocculateclays, the nonionic surfactants at first seem preferable. However, laboratory experiments should be performed using soil/aquifer samples and different types of surfac tants to determine potential flocculation problems. One successful pilot-scale implementation of surfactant-enhanced soil washing used nonionic and ionic surfactant washes in series. However, an alkali preflush was used to reduce the amount of exchangeable calcium and magnesium from the soil surfaces (9). In any event, the surfactant should be biodegradable and nonbiocidal so that it cannot harm the environment if it is injected and not recoverable. Soil/Aquifer System and Contaminant Interactions. Soil/aquifer systems and contaminants interact by many mechanisms as shown in Figure 2. Among these is the natural hydraulic gradient of the aquifer. This hydraulic gradient acts to push the N A P L through the soil matrix. The net result is a transport of N A P L away from the point of entry, thus enlarging the contamination zone. On the other hand, contaminant transport by hydraulic gradients can be used to remediate contaminated soils. Pumping wells can be installed down gradient from the contaminant source and used to induce an artificial hydraulic gradient


326


in the aquifer. Contaminants and water are then drawn toward the pumping wells and recovered. Capillary force is a second interaction mechanism which is measured quan titatively by the capillary pressure. Mathematically, it is the pressure differ ence between the N A P L and water phases, and is described using the Young & Laplace relation, equation 1 below (21). In writing this form of the Young & Laplace equation, we have assumed a hemispherical NAPL/water interface. This is the simplest case, but will allow easy estimation of the capillary pressure. pcap _ pNAPL

_ paq

_

^Jow

^

Γ Downloaded by FUDAN UNIV on March 15, 2017 | http://pubs.acs.org Publication Date: August 2, 1991 | doi: 10.1021/bk-1991-0468.ch016

cap

NAPL

aq

Here, P is the capillary pressure, p is the pressure in the NAPL, P is the pressure in the aqueous phase, η is the NAPL/water interfacial tension, and r is the radius of curvature of the hemispherical interface. Physically, the capillary pressure is a measure of the force necessary to de form a curved interface so as to allow it to pass through a constriction. The situation is illustrated in Figure 6 which shows a N A P L drop trapped in a pore. The natural hydraulic gradient in the aquifer contains insufficient energy to de form the interface at r and allow the droplet to pass through the constriction. Capillary forces are most significant for silt or clay-sized particles with a grain diameter of less than 0.075-mm. Equivalently, these are particles that will pass through a Tyler 200 mesh screen. The Young & Laplace equation can be used to estimate the order-ofmagnitude of the capillary forces that hold N A P L drops trapped in silt or claysized particles, and the hydraulic gradient required to push the N A P L drop through the constriction shown in Figure 6. Let P* and P% be the aqueous phase pressures on the up and down gradient sides of the droplet, respectively. Let r i and r be the radii of curvature of the interfaces as shown in Figure 6. The hydraulic gradient necessary to push the droplet through the constriction, ΔΡ, is then given by equation 2. 0%υ

2

q

q

2

AP = P? - P? = P

cap 2

Ρ

2

- Ρ Γ = Ίο„ (-

- -)

(2)

A typical assumed value for a NAPL/water interfacial tension is 3C dyne/cm. Further assuming a cylindrical pore geometry, and a hemispherical NAPL/water interface, r\ and r become the pore radii up and down gradients of the NAPL, respectively. Representative values for r\ and r may then be selected for clay-sized particles, and a few values representing the fine sands, the next larger particle size classification. Berg (22) provides a relationship for estimating pore throat diameters from particle diameters for rhombohedral packing of spheres which is also helpful in selecting useful η and r values. For example, π should vary from O-25-μπι and r from 0.5-10-μπι. Pore radii 2

2

2

2


16. CHAWLA ET A L


- NAPL SOLUBILIZATION - INTERFACIAL TNESION REDUCTION - CATIONIC, ANIONIC, NONIONIC - BIODEGRADABLE


- NONBIOCIDAL

Figure 5. Physical Properties of Surfactant

Figure 6. N A P L Drop Trapped in a Soil Pore Constriction


327

328


values less than 5.8-/xm correspond to silt or clay-sized particles while those between 5.8- and 32.3-//m correspond to fine-sand-sized particles. Using such values, the pressure gradients that are needed to push a N A P L drop one-cm in diameter through a given pore constriction can be estimated. Figure 7 shows the results in parametric form. Looking at the case where π is 5-//m and r is 0.5-/zm, a pressure drop of 10,800 kPa/m (480 psia/ft) is required to dislodge the N A P L droplet. This pressure drop is many orders of magnitude larger than the typical natural gradient of 0.13 kPa/m (0.006 psia/ft) that is ob served in aquifers which have no large elevation changes. This is certainly the worst case seen in Figure 6. However, even when r and π differ by as little as a factor of two, the required pressure drop is still very large compared to the typical natural gradients in aquifers. Thus, these calculations suggest that the natural gradients in aquifers are insufficient to mobilize NAPLs that are trapped as droplets in pore constrictions of silt or clay-sized particles. Similar calculations for fine-sand-size particles are also helpful as indicated by the curves for r at 5- and 10-μπι in Figure 7. In these cases, the pressure gra dients to dislodge a one-cm diameter drop were on the order of 0.1 to 0.2 kPa/m and are much closer to the natural gradients present in horizontal aquifers. Therefore, for particles in this size classification, the natural aquifer gradients may be sufficient to mobilize NAPLs that are trapped by capillary forces alone. The calculations just presented are only order-of-magnitude estimates. The real subsurface environment will have a pore size distribution. Variations of an order-of-magnitude in grain size are common between adjacent layers in an alluvial environment (72). Although simple, our calculations show how large capillary forces can be. In one sense capillary forces are good because they tend to hold the con taminant immobilized. On the other hand, many NAPLs are slightly water sol uble, a few ppm to a few hundred ppm. Therefore, these immobilized NAPLs dissolve into the ground water, in accord with their water solubilities, and are then transported along with the ground water. This phenomenon increases the size of the contaminated zone. To make the situation worse, some evidence suggests that the dissolution process is mass transfer limited (22). This means that dissolved N A P L concentrations in ground water are less than their solu bilities. The result is an increased dissolution time (i.e., time period required for a droplet to dissolve) and a greater long-term contamination problem for the ground water. Capillary and buoyancy forces influence contaminant movement in the ver tical direction in the soil/aquifer system. The N A P L can be thought of as an object partly immersed in a fluid. The fluid is air when the N A P L is in the vadose zone and the fluid is water when the N A P L is in the saturated zone. Archimedes' principle states that a body wholly or partly immersed in a fluid is buoyed up with a force equal to the weight of the fluid displaced by the body (25). In opposition to buoyancy is the gravity force that pulls the sub2


2

2




16. CHAWLA ET A L


329


330


merged object in a downward direction. The net force is given by the differ ence in buoyancy and gravity forces. This net force is directed either upward or downward in a vertical plane. If this net force causes the N A P L droplet to rise through the soil matrix, many pore constrictions will be encountered. At some point the N A P L drop may encounter a pore constriction that is very small. In this case the net upward force is much smaller than the capillary force necessary to push the drop through the pore constriction and the drop is trapped in the pore. The end result is an immobilized N A P L phase. A similar phenomenon occurs for NAPLs that have a net force directed in a downward direction; Berg provides a more general discussion of this interaction (22). In either case, the interaction of capillary and buoyancy forces results in N A P L emplacement in pores of the soil/aquifer system. Sorption represents another important interaction mechanism between the soil/aquifer system and contaminants. Depending on the forces of attraction, contaminant sorption to soil particles may be classified as physical, chemical, or electrostatic. Physical sorption is characterized by dipole-dipole interactions with relatively low heats of sorption in the range of 1-2 kcal/mole. Chemical sorption (chemisorption) represents the other extreme. It is characterized by covalent and hydrogen bonding with differential heats of sorption that can be as high as 50-100 kcal/mole. The last sorption category is electrostatic. Elec trostatic interactions are due to ion-ion and ion-dipole interactions. Two additional properties affect the soil/aquifer system-contaminant sorp tion interaction. These are the soil-cation exchange capacity and the surface area of the soil particles. Cation exchange capacity values of various soils con stituents are 2-6 meq/100 g for oxides and hydroxides, 80-150 meq/100 g for montmorillonite, and 200-400 meq/100 g for organic matter. Surface areas range from 7-30 m /g for kaolinite to 500-800 m /g for organic matter (24). From these values of cation exchange capacity and soil surface area, it appears that organic carbon content will have a larger effect on sorption of contami nants when compared to the clay mineral content. The sorption mechanism between the soil/aquifer system and the contam inant must be quantified in order to design effective in situ remedial action schemes. Simply put, it is necessary to estimate the amount of contaminant sorbed onto the soil particles. This sorbed contaminant will be more difficult to mobilize and recover when compared to contaminant that exists as a sepa rate nonaqueous phase liquid. A tremendous amount of work dealing with the sorption, transport and fate of all classes of organic compounds in subsurface soils has appeared in the liter ature (18,25-27). In particular, mathematical models have been used to corre late contaminant sorption with the aqueous phase contaminant concentration. Some of the more common mathematical models to predict sorption of con taminants onto soil particles are shown in Table I. These mathematical equa tions relate the equilibrium solid phase concentration, x/m (/xg of contaminant 2

2


16. CHAWLA ET A L


331

Table I. Sorption Models i/n

Freundlich Linear

x/m x/m

= KC = KC

Langmuir

x/m

= (x/m) K C/(l

p

P

max

+

p

K C) P


χ = mass of contaminant sorbed, μg m = mass of solid sorbent, g C = aqueous phase equilibrium concentration, mg/mL n = empirical constant, n > 1 sorbed per g of soil or sorbent), to the equilibrium liquid phase concentration, C (mg of contaminant/mL of liquid). The parameter K is the distribution coefficient. This coefficient is a mea sure of the strength of sorption and can be thermodynamically correlated to the Gibbs' free energy and the temperature. The Freundlich model is very general and fits most situations. In some situations it reduces to the simpler forms. It is linear if η = 1 and the Langmuir model results from the assumptions of constant differential heat of sorption and monolayer coverage. For hydropho bic organic compounds at low concentration, sorption isotherms (equilibrium plots of χ I m versus C at constant temperature) are linear. The value of K de pends on the properties of the contaminant and the soil. K is a strong function of the organic carbon fraction of the soil, f , and the octanol-water partition coefficient, K (26). K is a measure of the hydrophobicity of a compound. A normalized sorption coefficient is used in many sorption correlations. This normalized coefficient, K , is defined based on a soil that is 100% organic carbon. The mathematical definition is given in equation 3. p

p

p

oc

ow

ow

oc

K

M

= £

(3)

Joe 3

K varies widely for different contaminants (e.g., 1 cm /g for 2-propanol, 30,000 cm /g for oil, 200,000 cm /g for dioxin, and 5 χ 10 cm /g for grease) The sorption coefficient, K , is correlated with K for polynuclear aromat ics and chlorinated hydrocarbons (26), polynuclear aromatics (27), and halogenated alkenes and benzenes (18). These correlations are presented in Ta ble II. Their utility is that they can be used to estimate the amounts of contam inants that sorb onto soil particles. To estimate contaminant sorption, two physical parameters are needed: K and f . The K value for a contaminant can typically be found in the literature (12,28). The f value for a particular soil can be estimated or mea sured. With these values of K and f an appropriate correlation for K is selected from Table II and a K value is calculated from equation 3. Then an appropriate sorption model is selected from Table I and the amount of contam inant sorption is calculated. oc

3

3

6

p

ow

oc

3

ow

ow

oc

ow

oc

p


oc

332


Table Π. Koc Correlations

log K


oc

Application

Correlation

Ref

=

(.26) polynuclear aromatics & chlorinated hydrocarbons (27) polynuclear aromatics (18) halogenated alkenes & benzenes

\o K -0.21 ë

ow

= \o K 0.317 = 0 . 7 2 ^ / ^ + 0.49 ë

ow

The amount of contaminant sorbed onto soil particles is an important pa rameter in the design of any in situ remedial action plan. It is important to know what percentage of the contaminant will be sorbed on the soil particles versus the amount left either as a separate phase or dissolved in the ground water. Biodégradation is another interaction mechanism that may occur between contaminants and soil/aquifer systems. Microorganisms that are native to a soil/aquifer system, or that are acclimated to a contaminant and grow by natural selection and mutation, might be able to use the contaminant as a food source. If this happens, the contaminant will be transformed into a different species that may be more or less harmful than the original contaminant. The species that is formed depends on several variables such as: (1) the contaminant, (2) the microorganisms, (3) nutrient additions or availabilities, and (4) an aerobic or anaerobic environment. Biodégradation transforms the contaminant into a different species with different physical properties. In some cases, biodégradation can transform a molecule into a species that is nonbiodegradable (29). In other cases it can create a species that is also toxic. For example, one biodégradation pathway for trichloroethylene (TCE) produces vinyl chloride (30), another toxic sub stance. However, in the best cases of complete biodégradation under aerobic conditions, contaminants are transformed into carbon dioxide and water. NAPLs may also interact with soil/aquifer systems through the mechanisms of vaporization and diffusion in the vadose zone. Vaporization is due to the vapor-liquid phase equilibrium for NAPLs that are in contact with air. Vapor ized NAPLs diffuse through the vadose zone in the directions of decreasing concentrations. Also, the transport of vaporized NAPLs in a vadose zone will occur much faster than the corresponding N A P L transport in the water sat urated zone because the N A P L vapor phase diffusion coefficients are several orders-of-magnitude greater than the corresponding liquid phase diffusion co efficients. Typical vapor phase diffusion coefficients are on the order of 10" cm /s while those of the liquid phase are on the order of 10" cm /s (32). Thus, vaporization is a transport mechanism that increases the extent of the contam inated zone. 1

2

5

2


16. CHAWLA ET A L


333


Contaminants and soil/aquifer systems may also interact through plant root uptake. Plants absorb minerals through their roots and can also absorb contam inants in soils via this same mechanism. This causes contaminants to become part of the plant and results in their spread to the food chain. Finally, contaminants and soil/aquifer systems can react and form a differ ent species. Biodégradation falls into this category, but contaminants can also react with other species that may already be present in the soil/aquifer system (e.g., humic materials and other contaminants). The products of these reac tions are new species with different physical and chemical properties than the reactants. Their interactions with soil/aquifer system and with surfactants must then be investigated. Interaction Mechanisms between Surfactants and Contaminants. Surfac tants and contaminants interact through the mechanism of enhanced solubi lization. When a surfactant is added to water, it increases the aqueous phase solubilities of NAPLs. On the one hand, this increased solubility results in con taminated ground water. On the other hand, the contaminated ground water can be recovered using a recovery or pumping well. The recovered ground water can then be treated for contaminant removal; the cleaned water can be recharged to the aquifer. While this cleanup strategy has the potential to increase pollutant dispersion, it is important for many otherwise inaccessible sites. The surfactant mobilizes the contaminant so that it can be recovered. A surfactant does increase the aqueous phase solubilities of NAPLs, but dis solution rates are also important. Experimental evidence indicates that N A P L dissolution is mass transfer limited (32). However, Hunt et al. (12) present a model for N A P L dissolution kinetics which considers a spherical N A P L droplet suspended in a flowing aqueous stream and estimates the time required for the N A P L droplet to dissolve. Their model is based on a mass balance around a spherical N A P L droplet. The rate of change in the droplet mass is set equal to the rate of dissolution of the N A P L phase. The rate of dissolution is then expressed as a function of a mass transfer coefficient and the diameter of the N A P L drop. The resulting mass balance is an ordinary differential equation which can be solved to give the diameter of the drop as a function of time. The time at which the diameter goes to zero is defined as the droplet dissolution time. Relevant model parameters include the droplet diameter, N A P L solubil ity, and the flow velocity of the aqueous stream. This dissolution model was modified to estimate surfactant effects (e.g., in creased N A P L solubilities). For a given water flow velocity, the model dissolu tion times were computed for progressively increasing N A P L solubilities. The results are plotted in Figure 8 which illustrates the dissolution times for a spher ical, 10-cm droplet of TCE. Dissolution times for other size T C E droplets are given by Porzucek (11). The parametric curves in Figure 8 are for different values of the water flow


334



100-

U=0.01 cm/d U=0.1 cm/d U=1.0 cm/d U=10.0 cm/d U=100.0 cm/d

80 H

60

40 H

20

0.000

0.005

τ

0.010

0.015 3

AQUEOUS TCE SOLUBILITY (g/cm ) Figure 8. Trichloroethylene Droplet Lifetimes in an Aquifer as a Function of Aqueous Phase T C E Solubilities and Ground Water Flow Velocities (10 cm droplet). (Reprinted from Ref. 8.)


16. CHAWLA ET A L

Soil-ContaminantSurfactant Interactions

335

velocity, U. The top curve, U = 0.01 cm/day, shows dissolution times for a droplet in a very slowly moving aquifer. The curve for U = 0.1 cm/day (1.2 ft/year) is a realistic upper bound on naturally occurring ground water flow in the absence of large changes in aquifer elevation. The three remaining para metric curves are for larger ground water flow rates. These are more typical of aquifers that experience large changes in elevation or have an artificially in duced hydraulic gradient. The droplet lifetimes that correspond to a T C E solubility of 0.0012 g/cm are for systems without surfactant. As the surfactant concentration increases, the T C E solubility increases and the droplet lifetimes decrease. For the case where U = 0.1 cm/day, surfactant decreases the dissolution time from 30 years to approximatelyfiveyears and gives a quantitative measure of the magnitude of the surfactant effect. It also suggests that pump-and-treat remediation plans that don't use surfactants may require long-term strategies. Surfactants and contaminants can also interact by chemical reaction mech anisms. These mechanisms are not really desirable because they will keep a surfactant from performing its job of increased solubilization and lowering the interfacial tension. In general, laboratory studies should be used to eliminate candidate surfactants that could possibly react with contaminants. Surfactants and contaminants also interact through the mechanism of low ering interfacial tension. Surfactants that are injected into the ground water will eventually contact N A P L drops or lenses that are held in place by capillary forces. They will then lower the interfacial tension between the aqueous and nonaqueous (i.e., contaminant) phases. In common tertiary oil recovery exper iments, for example, surfactant formulations lowered the interfacial tension from approximately 35 dyne/cm to about 0.001 dyne/cm. Similar reductions are expected from surfactant-contaminant interactions although data are not now available. This decreased interfacial tension reduces the capillary forces to the point where it might be possible to dislodge the trapped N A P L drops by the natural pressure gradient in the soil/aquifer system. From the previous discussion, the pressure gradient needed for displacement of a one-cm N A P L droplet trapped in silt-sized particles was 10,800 kPa/m. Using surfactants, it may be possible to lower the required pressure drop to 0.1 kPa/m, a more re alizable value that could naturally exist in an aquifer. Once the N A P L drop is mobilized, it can be recovered using a strategically located pumping well. Surfactants and contaminants also interact by the mechanism of desorp tion since surfactants are capable of removing contaminants from soil parti cle surfaces. For strongly sorbed (i.e., chemisorbed) contaminants, however, a given surfactant may not be able to desorb all the contaminant from the soil. Thus, laboratory screening experiments for surfactants are essential to deter mine those formulations which are most effective. Another factor to consider, in addition to the desorption equilibrium, is the desorption rate. This rate is an important consideration because subsurface re-


3



336


mediation problems are expected to be mass transfer limited. These desorption rate experiments can be carried out in batch mode using doped soil samples, a surfactant solution, and a containment device such as a test tube. Doped soil samples are placed in a test tube and surfactant solution is added. The tubes are then agitated and allowed to settle. Samples of the aque ous solution are then drawn from the test tube and analyzed for contaminant concentration. Figure 9 shows results of typical batch experiments from our studies using TCE-doped soils and 0.5 vol% Triton X-100, a nonionic surfac tant. The soil used was a mixture of 80% Ottawa sand and 20% commercial top soil. Results are plotted as the cumulative percentage T C E desorbed as a function of time. Data are shown for three different TCE-contaminated soils. The first contaminated sample contained 12 μg TCE/g of soil while the second and third samples contained 25 and 34 μg TCE/g of soil, respectively. Equilib rium is reached in less than six hours and 80-95% of the original T C E on the soil is desorbed. Trichloroethylene is a slightly hydrophilic compound (up to 1200 mg/L dissolve in water) and it is not surprising that such high desorption rates and extent of desorption are achieved. A more hydrophobic compound on a soil with greater organic content is expected to show slower desorption rates and decreased percentages of contaminant removed. In such cases, the choice of surfactant is critical. Soil/Aquifer System and Surfactant Interactions. In a surfactant-enhanced, in situ soil-washing scheme, the hydraulic gradient of a pump is used to force flow aqueous surfactant solutions through the contaminated zone. Thus, the soil/aquifer system and surfactant interact because of the hydraulic gradient im posed by the pump. This hydraulic gradient not only causes flow of the aqueous surfactant solutions, it also creates mixing between the surfactant and contam inant phases. This mixing is desirable and hence larger hydraulic gradients are attractive. Unfortunately, large hydraulic gradients are not easily generated in soils because the soil matrix cannot withstand these pressures without fractur ing and losing integrity. Therefore, relatively low hydraulic gradients must be used for in situ soil-washing schemes, practical values are typically around 0.13 kPa/m. Surfactants dissolved in water will interact with the soil/aquifer system by sorption. By definition, surfactants are surface active species and like to sorb onto surfaces. More specifically, surfactants will sorb onto clay minerals and associate with dissolved humic materials that exist in soils. This sorption is influenced by the solution pH, cationic strength, and temperature. Finally, surfactants interact with the soil/aquifer systems through the mech anism of biodégradation. Native microorganisms present in a soil/aquifer sys tem may be capable of biodegrading a surfactant. Under aerobic conditions and complete mineralization, this process eventually converts the surfactant into CO2 and water. Moreover, it is desirable to use a biodegradable surfac-



16. CHAWLA ET A L


337

Figure 9. Cumulative Desorption from TCE-Contaminated Soil Contacted with 0.5 vol% Triton X-100 Surfactant



338

EMERGING TECHNOLOGIES IN HAZARDOUS WASTE MANAGEMENT Π

tant so that any surfactant left unaccounted for will simply degrade into harm less components. On the other hand, if the surfactant degrades too quickly, it will be ineffective at cleaning the soil. Thus, biodégradation tests should be performed using the selected surfactant and native microorganisms from the contaminated site. These tests will show how long a surfactant will survive be fore biodégradation starts. Several factors impede the application and flow of surfactant in a soil/aquifer system. Heterogeneities in soil type, porosity, and permeability will alter the fluid flow in soils and add uncertainty to the flow path of surfactant solutions. These factors make it difficult to obtain a uniform, predictable, areal sweep of injected surfactant solution through the desired soil/aquifer area. In addition to heterogeneities in soil physical properties, soil pore block age will also affect the flow of surfactant solutions. Such blockages can occur if emulsions orfloesform between surfactants and contaminants or surfactants and the soil/aquifer system. It should be possible to avoid such undesirable phase behavior by performing proper laboratory phase studies using the sur factants and contaminants. However, surfactant phase behavior is known to be a strong function of pH, temperature, cationic strength, and organic con centration (33,34). Unfortunately, these are variables that are not completely controllable at most in situ remediation sites. Therefore, it may not always be possible to avoid undesirable phase behavior and subsequent pore blockage. Pore blockage can also occur by mobilization of fines (e.g., clay-sized par ticles). Fines can be mobilized and redistributed by the natural hydraulic gra dient of the aquifer. Structurally, the fines consist of organic carbon and clay minerals and have a larger capacity for sorption when compared to other soil particles. Therefore, a large portion of a contaminant can sorb onto the fines. Once thefineshave redistributed in the soil/aquifer system, they create pref erential flow paths which bypass pores that are blocked by the fines. Surfactant solutions that have been injected into the soil/aquifer system will follow these preferential flow paths because they offer less resistance than the more tortu ous paths that go through the fines. When the fines segments of the soil are bypassed due to these preferential flow paths, the surfactant is ineffective in cleaning up the soil/aquifer system.

Summary Soil/aquifer system-contaminant-surfactant interactions are important in the design of in situ remedial action plans for contaminated sites that are either very large, very inaccessible, or both. Important physical properties of the soil/aquifer system, contaminant, and surfactant are identified and discussed as they pertain to further spread of contamination or cleanup. Interactions be tween these components and the influence of surfactants are also important.


16. CHAWLA ET A L

Soil-ContaminantSurfactant Interactions

339

The mechanisms of sorption/desorption, capillary forces, and mass transferlimited solubilization are particularly important.

Acknowledgements


This work was partially supported by Los Alamos National Laboratory under Contract No. 9-X58-8080U-1. We are thankful to Dr. Michael Ebinger of Los Alamos National Laboratory for his help in this study and Alexander Helou of Howard University for performing the laboratory studies.

Literature Cited 1. Ellis,W.,Payne, J., Tafuri, Α., and Freestone, F. In Proc Hazardous Mate rial Spills Conf, pages 116-124, Nashville, TN, 1984. 2. Ellis,W.,Payne, J., and McNabb, G. Treatment of Contaminated Soils with Aqueous Surfactants. Final Report EPA/600/2-85/129, Hazardous Waste Engineering Research Laboratory, Office of Research and Development, U.S. EPA, Cincinnati, OH, NTIS, Springfeld, VA, PB86-122561, 1986. 3. Chawla, R., Diallo, M., Cannon, J., Johnson, J., and Porzucek, C. Insitu treatment of soils contaminated with hazardous organic wastes using surfactants: a critical analysis. In Muralidhara, H., Ed., Solid/Liquid Sep aration: Waste Management and Productivity Enhancement, 1989 Int Sym, pages 355-367, Battelle Press, 1989. 4. Underground Movement of Gasoline on Groundwater and Enhanced Recov ery by Surfactants. API Publication 4317, Texas Research Insititute, 1979. 5. Test Results of Surfactant Enhanced Gasoline Recovery in a Large-Scale Model Aquifer. API Publication 4390, Texas Research Institute, 1982. 6. McNabb, Jr., G., Payne, J., Ellis, W., Kirstein, B., Evans, J., Harkins, P., and Rotunda, N . Chemical countermeasures application at the Volk Field site of opportunity. Internal Report to EPA, 1985. 7. Nash, J. Field studies of in-situ soil washing. Final Report to EPA on Contract No. 68-03-3203, 1986. 8. Nash, J. and Traver, R. In Proc Hazardous Material Spills Conference, Nashville, TN, 1984. 9. Sale, T. and Pitts, M . Chemically enhanced in situ soil washing. In Proc NWWA/API Conf on Petroleum Hydrocarbons and Organic Chemi cals in Groundwater: Prevention, Detection and Restoration, pages 487-503, Houston, TX, Nov. 15-17 1989. 10. Pouska, G., Trost, P., and Day, M. In Proc 6th National RCRA/Superfund conference and Exhibition, pages 423-430, New Orleans, LA, 1989.



340


11. Porzucek, C. Surfactant Flooding Technology for In Situ Cleanup of Contaminated Soils and Aquifers — A Feasibility Study. Report LA-11541-MS, Los Alamos National Laboratory, 1989. 12. Hunt, J., Sitar, N., and Udell, K. Nonaqueous phase liquid transport and cleanup 1. Analysis of mechanisms. Water Resources Res, 24(8):1247-58, 1988. 13. Wilson, J. and Conrad, S. Is physical displacement of residual hydrocar bons a realistic possibility in aquifer restoration? In Proc NWWA/API Conf Petroleum Hydrocarbons and Oranic Chemicals in Ground Water — Prevention, Detection and Restoration, pages 275-298, Houston, T X , Nov. 5-7 1984. 14. Hoag, G . and Marley, M . Gasoline residual saturation in unsaturated uniform aquifer materials. J Env Eng, 112(3):586-604, 1986. 15. Carter, C. and Suffet, I. Binding of D D T to dissolved humic materials. Env Sci Tech, 16(11):735-740, 1982. 16. West, C. Dissolved Organic Carbon Facilitated Transport of Neutral Organic Compounds in Subsurface Systems. PhD thesis, Rice University, Houston, TX, 1984. 17. Means, J., Wood, S., Hassett, J., and Banwart, W. Sorption of polynuclear aromatic hydrocarbons by sediments and soils. Env Sci Tech, 14(12):152529, 1980. 18. Schwarzenbach, R. and Westall, J. Transport of nonpolar organic com pounds from surface water to groundwater, laboratory sorption studies. Env Sci Tech, 15(11):1360-67, 1981. 19. Karickhoff, S. Organic pollutant sorption in aquatic systems. J Hydral Eng, 110(6):707-735, 1984. 20. Garbarini, D. and Lion, L. The influence of the nature of soil organics on the sorption of toluene and TCE. Env Sci Tech, 20(12):1263-69, 1986. 21. Adamson, A . Physical Chemistry of Surfaces. John Wiley & Sons, Inc., 1982. 22. Berg, R. Capillary pressures in stratigraphie traps. Am Ass Pet Geol Bull, 59(6):939-956, 1975. 23. Halliday, D. and Resnick, R. Fundamentals of Physics. John Wiley & Sons, Inc., 1974. 24. Bailey, G. and White, J. Factors influencing the adsorption, desorption and movement of pesticides in soils. Residue Rev, 32:29-92, 1970. 25. Weber, W. Sorption processes and their effects on contaminant fate and transport in subsurface systems. 1990 Lecture, Association of Environ mental Engineering Professors, 1990 Distinguished Lecture Series also submitted (with P.M. McGinley and L.E. Katz) to Water Res, Feb. 1990. 26. Karickhoff, S., Brown, D., and Scott, T. Sorption of hydrophobic pollu tants on natural sediments. Water Res, 13(3w):241-248, 1979. 27. Means, J. and Wood, S. Sorption of amino- and carboxy-substituted


16. CHAWLA ET AL.

28. 29.

30.


31. 32.

33. 34.


341

polynuclear aromatic hydrocarbons by sediments and soils. Env Sci Tech, 16(2):93-98, 1982. Leo, Α., Hansch, C., and Elkins, D. Partition coefficients and their uses. Chem Rev, 71(6), 1971. Karveta, L., Chung, H., Guin, K., Shebs, W., and Smith, L. Ultimate biodegradation of an alcohol ethoxylate and a nonylphenol ethoxylate un der realistic conditions. Annual Meeting Soap and Detergent Association, Boca Raton, Florida, Jan. 27-31 1982. Fisher, M. Transport and Fate of Organic Chemicals in Bandelier Tuff at Los Alamos National Laboratory Chemical Waste Site, New Mexico. Master's thesis, University of Texas, E l Paso, 1988. Bird, R., Stewart, W., and Lightfoot, E. Transport Phenomena. John Wiley & Sons, Inc., 1960. Sitar, N . , Hunt, J., and Udell, K. Movement of nonaqueous liquids in groundwater. In Proc Geotechnical Practice for Waste Disposal '87, pages 205-223, Geotechnical Division, American Society of Civil Engi neers, Ann Arbor, MI, 1987. Healy, R. and Reed, R. Physicochemical aspects of microemulsion flood ing. Soc Pet Eng J, 14(3):491-501, 1974. Healy, R., Reed, R., and Stenmark, D . Multiphase microemulsion sys tems. Soc Pet Eng J, 16(3): 147-160, 1976.

RECEIVED April 5, 1991


Importance of SoilContaminantSurfactant ... - ACS Publications

Recommend Documents